Computational & Technology Resources
an online resource for computational,
engineering & technology publications
engineering & technology publications
Civil-Comp Conferences
ISSN 2753-3239
ISSN 2753-3239
CCC: 11
PROCEEDINGS OF THE SEVENTH INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE, SOFT COMPUTING, MACHINE LEARNING AND OPTIMIZATION IN ENGINEERING
PROCEEDINGS OF THE SEVENTH INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE, SOFT COMPUTING, MACHINE LEARNING AND OPTIMIZATION IN ENGINEERING
Edited by: P. Iványi, J. Kruis and B.H.V. Topping
Paper 1.4
Enhancing Information Flow in Graph Neural Networks for Scientific Machine Learning
M. Chenaud1,2, J. Alves2 and F. Magoulès1
1MICS, CentraleSupélec, Université Paris-Saclay, Gif-sur-Yvette, Île-de-France, France
2, Transvalor S.A, Biot, France
Full Bibliographic Reference for this paper
M. Chenaud, J. Alves, F. Magoulès, "Enhancing Information Flow in Graph Neural Networks for Scientific Machine Learning", in P. Iványi, J. Kruis, B.H.V. Topping, (Editors), "Proceedings of the Seventh International Conference on
Artificial Intelligence, Soft Computing, Machine Learning and Optimization in Engineering", Civil-Comp Press, Edinburgh, UK,
Online volume: CCC 11, Paper 1.4, 2025,
Keywords: scientific machine learning, scientific computing, graph neural networks, Möller-Trumbore algorithm, inductive bias, physics-informed neural networks.
Abstract
The propagation of information through discretised domains is of crucial importance in the field of scientific machine learning. Recent studies have demonstrated the efficacy of graph-based models for physical simulations, particularly due to the inductive biases inherent in such frameworks. However, ensuring efficient information flow through these graph architectures is a delicate aspect, due to the wide range of scales of the simulated phenomena. We summarise some key architectural choices that are the most prominent in the literature, and we propose a novel edge augmentation technique, based on farthest point sampling and the Möller-Trumbore algorithm, for highly non-convex geometries. The efficacy of our approach is demonstrated through the training of a graph neural network model on a challenging, non-convex geometry.
download the full-text of this paper (PDF, 14 pages, 1057 Kb)
go to the previous paper
go to the next paper
return to the table of contents
return to the volume description